Implemetation And Application Of SHA-2 Hash Algorithms Based-on GPU

With the rapid improvement of GPU hardware manufacturing technology in the past more than one decade,GPU computing power increased at a rapid pace close to the Moore's Law.Due to the GPUs'programmability,high throughput and high concurrency capabilities,general-purpose computing on GPUs has become a research hotspot,and also increasingly being used for high performance computing.As one of two general-purpose computing on GPUs,the OpenCL platform offers a wider range of hardware support.This paper is based on OpenCL,studies rapid implementation of SHA-2 hash algorithms on GPU devices.Combined with the characteristics of GPU hardware structure,this paper proposed three GPU optimization principles.The principles are to reduce the computational complexity of GPU,reduce the overhead of GPU and increase the occupancy of GPU devices.Based on the principles,this paper summarizes a series of optimization methods of storage access and algorithm structure,and proposes a rapid implementation of SHA-2 series algorithm.In this paper,the rapid implementation of SHA-2 algorithms is applied to password recovery,and a high-performance GPU-based SHA-2 password recovery algorithm is designed and implemented.The algorithm supports multiple password recovery modes and implements parallel password hashing in different modes.The experiments show that the optimized SHA-2 password recovery algorithm effectively reduces the computational complexity and overhead of the algorithm,improves the device utilization and algorithm performance effectively,and has obvious advantages over the CPU and FPGA devices in password recovery.In mask mode,the optimized SHA-256 and SHA-512 password recovery algorithms achieve 1774 M and 390 M password hash/s,respectively,on AMD R9 290 GPU devices.